Stochastic Simulation of Selective Breeding Schemes for Wheat

We used stochastic simulation to investigate the effect of factors influencing long-term genetic gains in conventional wheat breeding schemes. In addition, we evaluated the long-term genetic gain using GEBVs (genomic estimated breeding values) as a selection criterion in a reference breeding scheme. The conventional wheat breeding schemes involve crossing parental lines and subsequent selfing of the offspring for several generations to obtain inbred lines. To understand the underlying factors influencing genetic gains, we simulated 16 breeding schemes using the Adamplant simulation tool. We used a finite-locus model to simulate these breeding schemes for 60 years. The breeding schemes consisted of different number of initial crosses (10, 50, 100, 120) performed every year which were made from 16 parents and variable number (200, 500, 1000, 2000) of single plants selected at the selfing stage F5. Phenotypic records were simulated in F3, F5, F6 and F7. Results showed that for each number of crosses, higher genetic gain was estimated with the highest (2000) single plants selected at F5, because higher genetic variance exists within a family than between families. The rate of genetic showed an increment by 21 and 83% when increasing the number of single individual from 200 to 2000 for the lower (10) and higher (120) number of crosses, respectively. When using GEBV as a selection criterion, we used genotype information from single F4 plants to do GEBV prediction from F5 to F7. Selection decisions were then made based on GEBV. The results will be included in the poster.